AREA 1. MEMBRANE PROTEINS IN BIOENERGETICS The research focus in Area 1 has been the ATP synthase, a membrane-bound enzyme that produces most of the cellular ATP. ATP synthases use a turbine-like rotary mechanism driven by the transmembrane flow protons down an electrochemical gradient, harnessed to power ATP production. Although key aspects of this process remain poorly understood, it is increasingly recognized that drugs that target the transmembrane domain of the enzyme might serve as potent antibiotics, as the ATP synthase is metabolically indispensable. One example is the FDA-approved drug Sirturo, used against drug-resistant strains of M. tuberculosis. Further progress in this area, however, will require development of compounds that are selective for the pathogens enzyme rather than that of the host. Thus, a focus of our recent studies has been the mechanism by which inhibitors of the membrane domain recognize the enzyme and impact its function. We have also continued to examine the role of the ATP synthase in other important physiological processes, aside from ATP production. In one study (Zhou W, Faraldo-Gmez J, BBA 2018) we studied how oligomycin binds and inhibits the mitochondrial ATP synthase. What makes this compound specifically interesting is that it does not inhibit any of bacterial enzymes, i.e. it demonstrates that this kind of inhibition can be highly specific. Yet little had been known about its mode of action at the molecular level. Using computer simulations, we showed that oligomycin naturally partitions into the lipid bilayer, preferentially residing at the lipid/water interface. We also showed that once in this this environment the inhibitor goes on to bind to membrane-exposed sites in the so-called rotor ring. These results strongly suggest that oligomycin physically blocks the rotary mechanism of this turbine-like enzyme. Importantly, the mode of interaction revealed by our simulations was corroborated by an experimental study of the structure of the enzyme in complex with oligomycin (Srivastava et al. Science. 2018). In a second study (Anselmi et al. JGP 2018), we utilized advanced simulation methods to substantiate a hypothesis we had previously put forward in regard to the role of the ATP synthase in defining the morphology of mitochondria (Davis et al, PNAS 2012). Specifically, we had proposed that mitochondrial ATP synthases self-organize into linear arrays comprising dozens of molecules, and that these supramolecular structures contribute to create the invaginations of the inner mitochondrial membrane known as cristae. In our follow-up study, we demonstrated that this process of self-organization is spontaneous and driven by the lipid bilayer; it is not activated by extrinsic factors, nor does it require direct interactions between adjacent proteins. Our studies showed that this driving force stems from the fact that isolated ATP synthases cause a pronounced deformation of the surrounding membrane, which is energetically costly. This deformation and the associated energetic penalty are reduced when the enzymes become arranged into linear arrays, with adjacent proteins side-by-side but not in contact. Importantly, this theory has been recently validated by an experimental study where purified mitochondrial ATP synthases were reconstituted into liposomes (W. Khlbrandt, personal communication); analysis of the spatial distribution of the enzymes using cryo-electron tomography proved that the hypothesized self-organization process takes place, and that it has a marked impact on the liposomes morphology. AREA 2. TRANSPORT AND SIGNALING ACROSS BIOLOGICAL MEMBRANES Two membrane protein families of major biomedical significance were the focus in Area 2. First, the family of cation/calcium antiporters, and specifically the Na+/Ca2+ (NCX) exchanger, which is key for the regulation of calcium signals, e.g. in cardiac cells controlling the heartbeat. Second, so-called MATE antiporters, which function as drug-efflux pumps and contribute to confer pathogens with multi-drug resistance. Both NCXs and MATEs are examples of secondary-active transporters. These proteins catalyze the transport of substrates across the membrane by coupling this process, directly or indirectly, to the translocation of a second species (typically Na+ or H+) down a pre-existing concentration gradient. In the NCX project, we continued our investigations of the mechanism of NCX_Mj, an archaebacterial member of the family. NCX_Mj is thought to be a valid model system of the cardiac Na+/Ca2+ exchanger, e.g. in future efforts in pharmacological research. Indeed, our previous computational studies have led to the prediction that the stoichiometry of the antiport cycle in NCX_Mj is 3Na+:1Ca2+, coinciding with that of the cardiac exchanger (Liao, Marinelli et al, NSMB 2016). Over the past two years we sought to evaluate this prediction through experimental functional studies, designed and conducted in collaboration with J. Mindell (NINDS). Specifically, NCX_Mj was purified and reconstituted in liposomes and its functional characteristics were evaluated through 45Ca2+ flux assays under a broad range of well-defined experimental conditions. Among other important findings, this study demonstrated that, as predicted, the transport stoichiometry of NCX_Mj is 3Na+:1Ca2+ (Shlosman et al., JGP 2018). These findings provide the foundation for future investigations of the mechanism of ion translocation. The project on prokaryotic MATEs was specifically focused on evaluating the structural basis for their specificity for either Na+ or H+ as the coupling ion (Ficici et al, PNAS 2018). MATEs are particularly intriguing in that some members of this family have been reported to be driven by both ions, synergistically, while others seem to feature a more conventional coupling mechanism. The underlying molecular basis for this variability is however largely unknown, in part because high-quality structural information is limited. To address this question we focused on PfMATE, from Pyrococcus furiosus. Analysis of available crystallographic data and additional molecular dynamics simulations unequivocally revealed previously unknown, highly selective binding sites for Na+ and H+, which are likely to explain the abovementioned functional observations. A bioinformatic analysis also revealed that these sites are broadly conserved. Importantly, an independent experimental study based on EPR spectroscopy (Claxton et al. PNAS 2018), published alongside our work, reached analogous conclusions for a close homolog of PfMATE, i.e. NorM-VC, confirming the location of the proposed binding sites. AREA 3. DEVELOPMENT OF SIMULATION METHODS The central achievement in Area 3 has been the development and release of a novel molecular-simulation method to facilitate the rigorous interpretation of experimental measurements techniques such as EPR or FRET (Hustedt, Marinelli et al., BJ 2018). This novel approach builds upon a method we had previously reported, known as Ensemble-Biased Metadynamics (Marinelli and Faraldo-Gmez, BJ 2016). In this method, a simulation is biased to sample a conformational ensemble that is exactly consistent with one or more probability distributions known a priori, thereby providing a faithful structural interpretation of the input data. The energetic cost associated with this bias can be quantified and used as a likelihood metric. In the new formulation, developed in collaboration with EPR experts at Vanderbilt University, the experimental errors of the target probability distributions are quantified, and accounted for explicitly in the simulations. This approach will contribute to set standards of quality and reproducibility in future studies of biomolecular dynamics through techniques such as EPR or FRET.